Urgent.News

What's breaking now, across thousands of outlets.

AI

Leaked video exposes how industrial-scale Chinese factories are dismantling RTX 5090s by the thousands to fuel AI servers

Wondering why RTX 5090s are so hard to find? Leaked footage shows thousands of consumer GPUs, including Founders Editions, being torn apart for AI servers.

Leaked video exposes how industrial-scale Chinese factories are dismantling RTX 5090s by the thousands to fuel AI servers

A leaked video has shed light on the alarming scale of Chinese industrial conversion factories dismantling GeForce RTX 5090 graphics cards in droves to fuel the growing AI server market. The footage, captured in various workshops scattered throughout China, showcases technicians meticulously desoldering the premium GB202 silicon and GDDR7 memory chips from brand-new RTX 5090 cards.

These components are then repurposed and installed into dual-slot blower design graphics cards tailored for server use. The conversion process is on an enormous scale, with thousands of RTX 5090 graphics cards being repurposed daily. Middlemen in the grey-market supply chain are constantly on the hunt for GeForce RTX 5090 stock to fuel these operations.

The GeForce RTX 5090, once a top-of-the-line gaming graphics card, has mysteriously vanished from retail shelves in China, where it was not even officially available. These industrial conversion factories are repurposing the cards to meet the high demand for AI compute clusters. The efficiency and scale of these operations are impressive, as the modified GeForce RTX 5090 graphics cards are then tested and repackaged for delivery to data centers or AI labs.

The high cost of the GeForce RTX 5090 has led to alternative solutions, such as prebuilt gaming systems, which sometimes offer better value than standalone graphics cards.

Written by urgent.news from Tom's Hardware's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.

Read the original at tomshardware.com →

More in AI

More from Saturday 24 October →